FOREIGN PORTFOLIO INVESTMENT AND

THE FINANCIAL CONSTRAINTS OF SMALL

FIRMS
By

(S.POTHINARAYANAN)

A PROJECT REPORT

Submitted to the

FACULTY OF INFORMATION AND COMMUNICATION ENGINEERING

In partial fulfillment of the requirements

For the award of the degree

Of

MASTER OF COMPUTER APPLICATION

IN
ANNA UNIVERSITY
CHENNAI 600 025

MAY, 2010

BONAFIDE CERTIFICATE
Certified that this project report titled…………………………is the bonafide work of
Mr./ Ms………………………………..who carried out the research under my supervision.
Certified further, that to the best of my knowledge the work reported herein does
not form part of any other project report or dissertation on the basis of which a
degree or award was conferred on an earlier occasion on this or any other
candidate.

Internal Guide The Head of the

Department

Internal Examiner External Examiner

 ABSTRACT

Title of Dissertation:
FOREIGN PORTFOLIO INVESTMENT AND
THE FINANCIAL CONSTRAINTS OF SMALL
FIRMS

This essay examines the impact of foreign portfolio investment on the

financial constraints of small firms. Utilizing a dataset of over 195,000 firm-
year observations across 53 countries, I examine the impact of foreign
portfolio investment on capital issuance and firm growth across countries
and firm characteristics, in particular size.

After controlling for firm-, industry- and country-level characteristics such as

change in foreign exchange rate, share of market capitalization, relative
interest rates and investment climate, I find that foreign portfolio investment
helps to bridge the gap between the amounts of financing small firms require
and that which they can access through the capital markets.

Specifically, I find that foreign portfolio investment is associated with an

increased ability to issue publicly traded securities for small firms in all
nations, regardless of property rights development. For those small firms
that do issue, the form of capital appears to be debt. Since small firms often
rely heavily on bank lending, I also test for potential increases in credit for
small firms utilizing the bank lending theory of monetary transmission.

Results show significantly decreased short-term debt and increased long-

term debt, supporting the contention that bank debt maturity to these firms
has increased. This transition to longer-term debt could also be as a result of
the increased public debt securities these firms are more able to access.

The overall increased access to capital only leads to value-enhancing growth

at the firm level in nations with more developed property rights. I find that
the volatility of foreign portfolio investment is significantly negatively
associated with the probability of small firms
Issuing publicly-traded securities as well as their firm growth, in periods
when their domicile nations are deemed less ‘creditworthy.’ Results
underscore the significance of a good financial system that minimizes
information asymmetry and enhances liquidity, as well as property rights and
country creditworthiness, to instill confidence in foreign investors.

S.POTHINARAYANAN

MBA (FINAL YEAR)

Introduction
“The causes of the currency crises in emerging markets during the late
1990s
have been the subject of much debate—especially considering that, before
the
crises, many of the Asian countries tended to have balanced budgets and
generally sound macroeconomic performance. …Some observers argue that
given
the generally favorable macroeconomic conditions, that the crises were not
caused by incompatibility between fiscal and monetary policies and
exchange rate
pegs, but rather by the unexpected and self-fulfilling panics of foreign
investors.”
Federal Reserve Bank San Francisco: Economic Letter
Although most policy-makers encourage the opening of financial markets to
foreigners, foreign portfolio investment, or ‘hot money,’ seems to be
perceived in a very
negative light. Its short-term nature seems to be blamed for almost every
economic ill
when it comes to crises and to well-publicized disputes such as that between
former
Malaysian Prime Minister Mahathir bin Mohammed and financier George
Soros. It is
further often compared to its much more stable and longer-term global
capital flow
counterpart, foreign direct investment and found lacking. If countries have
opened their
borders to foreign investors and maintained a dearth of capital controls, it
would seem
intuitive that there exist at least some positive attributes to this form of
private capital
flow. Indeed, a publication of the International Finance Corporation states the
following:
“In many markets, relatively small amounts of foreign capital have been
enough to act as a catalyst. …Foreign [capital] entry typically has set off two
parallel virtuous cycles. First, an institutional development cycle.
International
investors—and, by proxy, the managers of their funds—demand high
standards in
regulation and information. Spurred by the prospect of new investment,
market
regulators often undertake reform as part of the opening-up process. In
addition,
fund managers require local services, such as brokers, custody and transfer
agents, and information on local companies. In response to this demand,
local
providers spring up, competition increases and standards improve.”
Viewpoint, The World Bank Group
2
In this paper, I strive to understand better the impact of foreign portfolio
investment and whether there exists a positive impact of foreign portfolio
investment.
Specifically, in the first chapter, I examine whether foreign portfolio
investment, or ‘hot
money,’ eases the financial constraints of small firms. I further examine
which route this
impact takes: a direct route, through the capital markets, or an indirect
route, through
bank lending. Finally, I examine whether this impact ultimately leads to
growth at the
firm level. Due to the information asymmetry and agency costs associated
with foreign
portfolio investment (also referred to as FPI), it isn’t immediately obvious
whether small
firms would be able to access this additional source of financing. In the
second chapter, I
examine how the instability of foreign portfolio investment impacts the
financial
constraints and whether this impact destroys the benefits of foreign portfolio
investment
with regard to access to finance and ultimately, the growth of small firms.
Given the recent emphasis in literature on global cash flows and access to
finance,
this study may find an audience with those that follow this area of financial
literature
such as advocates of liberalization and policy makers. Small firms that have
access to the
public capital markets, as well as organizations whose intent it is to support
small firm
vitality may also be interested.
I find that financial constraints are eased by foreign portfolio investment for
small
firms in nations with more developed property rights (also referred to as DPR
nations), as
demonstrated by an increase in the probability that such a firm in need of
financing issues
capital when there is an increase in foreign portfolio investment in its
domicile nation. I
also find evidence that small firms in nations where property rights are less
developed
(LDPR nations) are helped by foreign portfolio investment as demonstrated
by a
3
significant increase in the probability of issuing capital. Conditional on a firm
reaching
the capital markets, the probability of equity issuance is negatively
associated with an
increase in foreign portfolio investment for small firms. These results,
coupled with
evidence of an increased access to capital provided by foreign portfolio
investment,
imply that the form of capital for the small firms that access it is debt. The
route that
foreign portfolio investment takes to ease the financing constraints of small
firms in DPR
nations is primarily direct, as seen through the increased probability of
capital issuance
and only very weak evidence of an indirect route, as seen in a modest
increase in the
liquidity of bank balance sheets combined with a decrease in the level of
domestic credit
provided. The positive impact of foreign portfolio investment on access to
finance, either
directly or indirectly, translates into growth for small firms in only those
nations with
more developed property rights (DPR), which underscores the importance of
an attractive
investment climate and growth-nurturing financial development in a firm’s
domicile
nation.
With regard to FPI volatility, I do not find evidence that the volatility of
foreign
portfolio investment (also referred to as FPI) - examined without
distinguishing terms of
relative confidence in a country (or country “creditworthiness”), as measured
by a rating
by institutional investors - is damaging to small firm access to finance, as
measured by
the probability of public capital issuance. This is true regardless of the
development of
property rights. Once examined in subsets of “investment grade” and
“noninvestment
grade” country years, a proxy for the level of confidence institutional
investors have in a
country’s distance from crisis, and controlling for the level of foreign
investment, I find
that small firm access to finance in countries is only significantly negatively
associated
4
with the volatility of FPI in the noninvestment grade subset. Importantly, I
also find that
the volatility of foreign portfolio investment only hinders the growth of small
firms only
when nations are deemed less ‘creditworthy’ or closer to crisis. These results
imply that
the benefits derived from FPI such as increased liquidity and an enhanced
investment
environment (Levine and Zervos (1996)) through better corporate
governance, investor
protection and transparency (Feldman and Kumar (1995)), which have been
linked to an
increased level of access to finance (La Porta, Lopez-de-Silanes, Shleifer and
Vishney
(2000) – henceforth these authors are referred to as LLSV) and to more
efficient
allocation of capital (Wurgler (2000); Love (2001); Rajan and Zingales
(1998)), are not
depleted in times when country risk is deemed low enough in nations where
investors are
protected. This finding is striking given the fact that volatility is actually
larger on
average in those countries considered investment grade.
The papers most related to mine are Harrison, McMillan and Love (2004)
(henceforth these authors are referred to as HML) and Laeven (2003).
Although similar
in intent, these papers differ from mine on many dimensions. HML (2004)
focuses on
the impact of foreign direct investment1 and examines this capital flow as a
proportion of
all investment, controlling for the proportion of total investment rather than
the size of
the market in question. Laeven (2003) examines the impact of liberalization
or reform
policies on financial constraints rather than the specific cash flows resulting
from said
reformation. These papers also differ from mine with regard to methodology.
Both
1 Foreign direct investment is defined by the IMF (Balance of Payments
Manual (1993)) as inflows of
investment including short-term and long-term equity capital and
reinvestment of earnings for the purpose
of acquiring a lasting management interest in a foreign company. Foreign
portfolio investment is defined
by the IMF (Balance of Payments Manual, 1993) as equity and debt issuances
including country funds,
depository receipts and direct purchases by foreign investors of less than
10% control.
5
papers focus on the result of financial constraints (i.e. investment sensitivity
to cash)
whereas I look to the source of the financial constraints, capital markets and
bank credit.
Utilizing the Euler equation, these papers examine implied external financial
constraints
instead of more direct evidence - capital issuance, or the lack thereof.
External financing
(i.e. issuing equity and debt) can be used to finance investments and is in
fact used when
firms are financially constrained. As such, evidence of this form of financing
should be
considered when examining financial constraints (Fazzari, Hubbard and
Petersen (1988)).
In fact, the type of security issued can be indicative of the level of financial
constraint a
firm faces (Chittenden, Hall and Hutchinson (1996) - henceforth these
authors are
referred to as CHH). Importantly, these papers differ from mine with regard
to their data.
Both HML (2004) and Laeven (2003) utilize the Worldscope database.
Inasmuch as my
emphasis is on small firms, I create a unique database of over 195,000
observations and
across 53 countries to circumvent the larger firm bias from which many
existing
international databases suffer. Given the considerable differences across
focus,
methodology and scope, it is difficult to compare results of these papers with
mine.
HML (2004) find that foreign direct investment as a proportion of overall
investment
alleviates financial constraints and that foreign portfolio investment as a
proportion of
overall investment does not. Although arguably these results are in
contradiction with
my results superficially, the differences in small firm focus, data and
definition of foreign
portfolio investment leave room for differences in these results and
consequently are not
directly challenged by my results. Laeven (2003)2 finds that liberalization
alleviates
financing constraints. Although the scope of his research question is
significantly
2 Samak and Helmy (2000) provide a very thorough analysis of foreign
portfolio investment in Egypt but is
not a true empirical work so is not considered a related paper.
6
different, my results are not in disagreement of those of Laeven (2003)
inasmuch as I
provide results specific to property rights development subsamples, which
can be loosely
compared and whose results provide additional support to those of Laeven
(2003).
This paper contributes to three main areas of literature. The first is small firm
access to capital. As markets become more integrated, foreign portfolio
investment is a
potential source of new investment capital for these financially constrained
(Beck,
Demirgüç-Kunt and Maksimovic (2005) – henceforth these authors are
referred to as
BDM) firms. Information as to whether this additional source of capital for
small firms is
feasible given the information and agency environments is useful in
extending this
literature.
This work is also related to the literature on global capital flows. As more and
more countries consider reforming foreign investment policy to enable
capital market
integration, this area of research becomes a resource for many.
Lastly, this research touches on that of liberalization. Although, not a study
on
liberalization, this paper offers insight into the impact of one potential factor
in a
country’s investment environment, foreign portfolio investment. Capital
market
liberalization opens country borders to foreign investment, which may
ultimately broaden
and deepen financial markets but can also open countries to vulnerability to
the fickleness
of foreign investment. Understanding what drives the aftermath of
liberalization, such as
the impact of a change in foreign portfolio investment, may offer insight into
the debate
on liberalization.
7
Chapter 1: Can Foreign Portfolio Investment Bridge the Small Firm
Financing Gap Around the World?
I. Motivation
A. The Challenges of Small Firms
According to a report of the President on the state of small business for the
year
1999-2000, small businesses3 represent 99% of all businesses, employ half
of those
Americans who are gainfully employed and create two-thirds of the job
openings that
occur in the United States. Other countries have a similarly large proportion
of total
firms represented by the small firm (Klapper, Sarria-Allende and Sulla
(2002)). These
firms often face a lack of liquidity (BDM (2002); Beck, Demirgüç-Kunt, Laeven
and
Maksimovic (2003)) and an excessive sensitivity to government regulation
that could
result in premature failure (BDM (2005); Tewari and Goebel, (2002)).
Accessing public
equity capital is generally even more difficult for small firms (CHH (1996)).
When it can
be, it is more expensive for the small firm relative to large firms (Warner
(1977); Smith
(1977)). The size-bias small firms face from potential investors is one of the
main
challenges facing these firms. The cumulative effect of these challenges is
called a
“finance gap” (Macmillan Committee (1931); Wilson Committee (1979)) and
reflects the
lack of capital available to these informationally opaque firms (Berger and
Udell (1998)).
3 Small firms here are defined by the Small Business Administration as any
firm that does not exceed at
least one of the following: 1) 500 employees for most manufacturing and
mining industries, 2)
100 employees for all wholesale trade industries, 3) $6 million for most retail
and service industries, 4)
$28.5 million for most general & heavy construction industries, 5) $12 million
for all special trade
contractors or 6) $0.75 million for most agricultural industries
8
Looking to the international capital markets for funding does not necessarily
make matters much better for the small firm. Beyond the same biases facing
them in the
domestic market, small firms face challenges from international capital
sources for
several reasons – most notably, supply. Increasing the supply of capital,
either
domestically or internationally, could at least in part remedy this situation.
HML (2004) find that global capital flows are associated with a reduction in
the
financing constraints of firms. As an increasingly vital part of global capital
flows (see
figure 1), foreign portfolio investment helps to increase financial
development thus
furthering the cause of decreasing financing constraints (Love (2003)) and
more
specifically, enhancing the development of markets (Levine and Zervos
(1996)), which
affords firms more opportunities to raise external capital (Demirgüç-Kunt and
Maksimovic (1998)). Perhaps most important for this analysis is the fact that
the
competition for these cash flows motivates an improvement at the firm level
in such
things as transparency, disclosure and corporate governance (Bekaert and
Harvey (2003);
Evans (2002); Levine and Zervos (1996); Feldman and Kumar (1995)), which
is
particularly important for those investors looking to invest internationally
(Aggarwall,
Klapper and Wysocki (2003)). The financially-constrained small firms striving
to
capture some of this additional source of capital have an even greater
incentive to do so.
B. The Path of Investment
Small firms tend to be partially dependent on bank lending to finance their
growth
(Warner (1977); Cull, Davis, Lamoreaux and Rosenthal (2004) – henceforth
these
authors are referred to as CDLR). A portion of small firms reaches the capital
markets
and may use public financing to grow their businesses. Those who do go
public may still
9
rely at least in part on bank lending to finance their growth. As such, there
are two ways
in which foreign portfolio investment may reach the small firm: the ‘direct’
route -
through the capital markets, and the ‘indirect’ route - through banks who will
in turn
invest in or be able to extend more credit to these small firms (see Figure 2).
B.1 The ‘Direct’ Path of Investment
If competition for a scarce resource, such as capital, improves the
investment
environment through superior transparency, disclosure and/or corporate
governance – an
effective improvement in the investment environment, the set of firms in
which foreign
investors consider investing is increased to include some firms who
previously had
difficulty obtaining financing due to information asymmetry and/or agency
costs. This
implies an improvement in the allocation of capital which has been
associated with
market development (Wurgler (2000)). Small firms with their informationally
opaque
nature may be included in this marginal group of firms. This is relevant due
to the
challenge of small firms in accessing capital in any form. Firms perceived as
‘investible’
who need external financing should realize an increased probability of
domestic capital
issuance with an accompanying increase in foreign portfolio investment. An
increase in
the probability of capital issuance stems from the increase in supply of
capital and is not
identified as due to foreign or domestic investors. I examine whether the
level of foreign
portfolio investment helps to ease the financial constraints of small firms.
More
explicitly stated,
H1a) The probability of capital issuance for small firms is significantly
positively
related to the level of foreign portfolio investment of a country (e.g. the
financial
constraints of small firms are relaxed).
10
Beyond whether a firm issues, I examine the type of security a firm issues.
Inasmuch as small firms are typically debt-laden (CDLR (2004)), the ability to
issue
equity could be perceived to be a greater alleviation of financing constraints
since there
are no fixed payments associated with this form of capital. This ‘choice’ of
capital form,
therefore, becomes informative. Not much has been written in the
international arena
examining the feasibility of capital choice for constrained firms. Korajczyk
and Levy
(2003) provide an examination of capital structure choice for both financially
constrained
and financially unconstrained firms in the United States. Although solely a
domestic
study, the main result in Korajczyk and Levy’s work is that constrained firms
issue what
they can when they are able. There isn’t any compelling reason, beyond an
increased
disclosure and governance at the firm-level, that would lead us to believe
that these firms
would be able to access equity as a result of the increase of the supply of
capital available
to firms, domestically or internationally. We should see that although small
firms will
indeed see an easing of their financial constraints, this easing would mainly
be in the
form of debt capital4. To that end, I hypothesize the following:
H1b) Conditional on firms issuing capital, the probability of small firms
issuing
equity capital will not be significantly positively related to foreign portfolio
investment.
B.2 The ‘Indirect’ Path of Investment
For those firms who are dependent on bank lending and/or remain unable to
access publicly issued securities, the ‘direct’ path of foreign portfolio
investment is
irrelevant. An ‘indirect’ path through financial institutions instead is relevant.
The
theory behind this path of investment stems from the bank-lending theory of
monetary
4 Also see Henderson, Jegadeesh, and Weisbach (2004).
11
policy (Bernanke and Blinder (1988); Kashyap and Stein (1995); Kashyap,
Rajan and
Stein (2002)). Kashyap and Stein (2000) is particularly relevant in that they
find that
small banks are particularly sensitive to monetary policy. This is relevant
since small
banks are most likely to be the banks to serve small firms. The lending
theory finds that
money supply tightening (expansion) appears to decrease (increase) the
ability of banks
to loan funds based on the relative illiquidity of their balance sheets. What
this implies is
that if there is a positive money shock into a country, bank balance sheets
become
relatively more liquid thus enabling them to increase the amount of credit
extended to the
public. Although this money supply augmentation is due to monetary policy
in Kashyap
and Stein’s paper, this theory could be extended to consider a different
source of ‘money
supply’ – in this case foreign portfolio inflows. An increase in the liquidity of
the bank’s
balance sheet through increased outside investment enables banks to lend
in the same
manner as if there were a change in money supply caused by monetary
policy5. More
concisely stated,
H2) The liquidity of bank balance sheets, as well as the amount of domestic
credit, are significantly positively related to the level of foreign portfolio
investment of a country.
C. Foreign Portfolio Investment as a Derivative of Growth
Although it is informative to know whether foreign portfolio investment
increases
small firm access to finance, it is at least equally important to know whether
that
additional access to finance helps the firm to grow. Assuming the investment
environment is sufficient to nurture small firm growth and foreign investment
in general
(as is the case in nations with developed property rights); we would expect
this to be the
5 This can be through investment in the banking sector or due to the
implications increased money supply
12
case. In other words, if the property rights of a nation support investment
sufficiently
that small firms may realize less financial constraint; the additional access of
capital
provided by FPI should translate into firm growth. According to Guiso,
Sapienza and
Zingales (2003), the integration of capital markets leads to local financial
development,
and ultimately to small firm growth. In this context, it could be inferred that
to the extent
that an increase in foreign portfolio investment coincides with the integration
of capital
markets we would expect the small firms to grow at a rate closer to their
unconstrained
growth rate. To examine whether that in fact is true, I test whether foreign
portfolio
investment helps small firms to grow.
Empirically tested this becomes,
H3) The growth rate of small firms, as defined by the percentage change in
total
assets (and separately sales revenue), is significantly positively related to
the
level of foreign portfolio investment of a country in DPR nations.
II. Methodology
A. To Issue or not to issue
Inasmuch as my goal is to examine the impact of FPI on small firm financing
constraints, I utilize the findings of BDM (2005) and assume that all small
firms are
financially constrained. As such, I limit my sample to small firms, as defined
by the
bottom tercile of firms ranked by total assets6. In so doing, I assume that
any firm-year
where capital issuance does not occur represents financial constraint.
The results of a Durbin-Wu-Hausman test for endogeneity for foreign
portfolio
investment demonstrate that endogeneity is a concern. In support of this
evidence are the
has on the ability of banks to raise reservable forms of finance.
6 Size terciles are created annually within countries so that firms are allowed
to move into and out of size
categories.
13
results of Agarwal (1997), which shows that the significant determinants of
foreign
portfolio investment are the (change in the) real exchange rate, share of the
domestic
capital market in the world capital market and some proxy for economic
activity. As
such, I utilize an instrumental variable methodology that in the first stage
regresses total
foreign portfolio investment (average across years t-3 through t-1) on the
above variables
with relative interest rates (country interest rates scaled by world interest
rates) serving to
explain the economic activity and add to this list liberalization and FPI
volatility to
predict foreign portfolio investment with its significant determinants. The
reason behind
the additional variables is based on much of the liberalization work down by
Bekaert and
Harvey, Henry, and Patro and Wald7. The first stage regression then
becomes:
FPIj,t = ?0+ ?1AFXRatej,t + ?2Sharej,t + ?3RelIntRatesj,t + ?4Libj,t +
?5FPIVolj,t + t + E (1)
To discern the impact of foreign investment on access to capital (as proxied
by the
probability of capital issuance), I divide the sample into halves based on
property rights.
Controlling for country and industry fixed effects, as well as firm-level capital
structure
choice determinants (such as cash flow, debt/asset level, profitability, risk,
growth, and
asset tangibility) to ensure that my results determine the extent of the
access firms
possess to issue necessary external capital (versus discretionary choice), I
isolate the
impact of foreign portfolio investment on the ability of firms to access
capital. Why the
“potentially”? If it is not significant in DPR countries you would not expect it
to be
significant n LDPR countries. If it is that calls for more set-up at this stage.
Empirically, I
use an instrumental variable probit model wherein I first instrument the
foreign portfolio
7 See Bekaert and Harvey (2003), Henry (2000; 2003) and Patro and Wald
(2004).
14
investment flows (scaled by gross domestic product) and in a second stage
estimate the
impact of these flows on the ability of small firms to access capital. The
probit
methodology is used due to the limited nature of the dependent variable and
the panel
format of the data. The dependent variable in this model is the probability of
capital
issuance (i.e. occurrence of issue: y = 1; no capital issuance: y = 0).
Prob(y=1)i,t = F0+ F1FPIj,t-1 + F 2 Yj,t-1 + F 3 Xi,t-1 + Ii+ t + E (2)
where FPI is the predicted level of foreign portfolio investment from the first-
stage
regression (represented in equation (1)) in the instrumental variable probit
regression (see
equation (1) for the first stage), X is a vector of lagged firm-specific variables
such as
cash flow, debt/asset level, profitability, risk, growth, and asset tangibility.
These
variables control for occurrences wherein firms would be more likely to issue
(Korajczyk
and Levy (2003)). Y is a vector of lagged macroeconomic variables such as
GDP
growth, levels of other potential sources of capital such as foreign direct
investment and
savings, and variables of development such as private credit, law and order
and
corruption levels. Macroeconomic variables are calculated as three year
trailing moving
averages in order to abstract from business cycles. This methodology is
often used in
cross-country analyses to smooth out annual fluctuations that can otherwise
confound
results (see Beck, Demirgüç-Kunt and Levine (2003) – henceforth these
authors are
referred to as BDL; Rousseau and Wachtel (2002)8). I is a vector of industry
dummies to
control for industry effects and t represents time dummies, which control for
any time
effect in the panel. A description of the firm-, industry- and country-specific
variables is
8 See also Beck, Demirgüç-Kunt, Laeven and Maksimovic (2003).
15
in the data section as well as in the appendices9. The instrumental probit
methodology
utilizes weights frequency to avoid data cloning issues and a bootstrapping
methodology
which uses randomly chosen subsamples10 of the dataset with replacement
to avoid
dependence on assumption of the normality of distribution or the absence of
stochastic
influences on the data. The bootstrapping technique is vital in ensuring that
the standard
errors are correct and that resulting significance is accurate. Based on my
hypothesis,
H1, I expect T1 to be both positive and significant for small firms in DPR
countries and
potentially for small firms in LDPR countries.
B. Capital Choice
The relevance of capital choice lies in its informative nature with regard to
the
extent of the relaxation of a firm’s financial constraints. Due to the reliance
of small
firms on debt, the issuance of equity capital could be perceived as an
increased easing of
financial constraints.
Inasmuch as the debt vs. equity capital “choice” implies that the capital
issuance
variable is positive (y=1 in equation (2)), equation (3) represents the second
stage of the
conditional logit model begun in the previous section. This model enables me
to estimate
the extent to which foreign portfolio investment alleviates financial
constraints, since
enabling small firms to obtain equity capital would not only meet their need
for capital
but also provide much needed financial flexibility by offering this capital
without the
fixed payments associated with debt. In the model illustrated below, the
dependent
variable represents capital choice, specifically y=1 for equity issuance and
y=0 for debt.
9 Tobin’s Q is not included in my analysis due to the sparsity and lack of
consistency of information on
market pricing in both less and more developed nations around the world.
10 N=50 is used for bootstrap replication.
16
Prob(y=1)i,t = H 0+ H1FPIj,t-1 + H 2 Xi,t-1 + H 3 Yj,t-1 + Ii+ t + E (3)
where definitions of variables are as in equation (2). Once again controlling
for country,
industry and time fixed effects as well as frequency of country observations
and utilizing
bootstrapping techniques, I ascertain the affect of foreign portfolio
investment on the
ability of small firms to issue longer-term capital. Given that the vast
majority of the
additional capital coming into the financial markets is debt (Henderson,
Jegadeesh and
Weisbach (2004)) and the fact that these firms typically suffer from an
inability to access
equity capital (CHH (1996) ; CDLR (2004)), domestic or otherwise, I see no
compelling
reason that the probability of issuing equity capital would be significantly
increased. I
anticipate that this coefficient, ��� will not be significant for small firms
since these firms
typically suffer from an inability to access equity capital, domestic or
otherwise.
C. Domestic Credit
To address those small firms in my dataset that are at least in part reliant on
bank
debt, I examine the impact of foreign portfolio investment on domestic
credit. I examine
the impact of foreign portfolio investment separately on two proxies of credit
availability
as well as a measure of bank balance sheet liquidity. Utilizing a cross-
sectional timeseries
of country-level data, I first regress the following:
Debtj,t = I0 + I1FPIj,t-1 + I2Yj,t-1 + t + E (4)
where Debtj,t represents two proxies (regressed separately) for the level of
credit available
in country j. These proxies include the level of domestic credit available and
the level of
domestic credit available that is provided by banks. If the implications of the
banklending
channel are true in the case of foreign portfolio investment, the coefficient
on
17
FPI, I1, will be positive. Clustering at the country level is undertaken to avoid
any issues
of data cloning.
Utilizing the same equation, I test the potential impact of foreign portfolio
investment on changes in the bank balance sheet liquidity by utilizing a
proxy for bank
liquidity – Bank’s liquid reserves to asset ratio. Empirically, this becomes the
following:
Liquidityj,t J0 + J1FPIj,t-1 + J2Yj,t-1 + t + E (5)
Other definitions of variables are once again as they are in equation (2).
Additional
variables included in Yj,t-1 are fiscal burden, based on the tax implications of
credit
(Desai, Foley and Hines (2004)), and relative interest rates, based on the
cost of debt
implications (Kashyap and Stein (2000)) on the demand for bank credit. If
foreign
portfolio investment does increase the liquidity of the balance sheets of
banks as the
theory implies, we would expect to see a positive J1.
I also test the impact of FPI on bank credit at the firm level. This provides an
opportunity to examine how not only total leverage changes with levels of
FPI, but also
the maturity of the debt utilized by small firms. Indeed, a decrease in the
level of shortterm
debt, a debt maturity upon which these firms most typically depend (Barclay
and
Smith (1995)), in favor of longer-term debt, i.e. an increase in the maturity of
outstanding
debt, would imply a reduction in financial constraints as longer-term debt
involves less
interest rate risk and provides capital over a longer term. Utilizing a similar
version of
the regression in equation (2) that uses as its regressand leverage I regress
the following:
(Short-term/Long-term/Total) Lev i,t = L0+ L1FPIj,t-1 + L2Yj,t-1 + L3 Xi,t-1 +
Ii+ t + E (6)
18
where Levi,t refers to the amount of leverage (of each type separately) firm i
holds in time
t and all other variables are as they appear in equation (2). Additional
variables real
interest rates and fiscal burden are used to control for capital choice as well
as overall
credit demand/supply issues. L1 for short-term debt will be positive if access
to capital
has increased enough to reduce small firm dependence on bank lending. If
the contention
that FPI helps to alleviate the financial constraints of small firms through an
extension of
the maturity of their outstanding debt is true, the coefficient on FPI in
specifications
using long-term and total debt as the regressand, L1, will be. Frequency
weights and
bootstrapping techniques are once again utilized.
D. Growth
To examine whether the direct or indirect route of foreign portfolio
investment
ultimately leads to firm growth, I utilize the growth rates of these firms,
ascertaining
whether foreign portfolio investment impacts their growth (both in sales and
in total
assets) by regressing the following:
Growthi,t = N0+ N1FPIj,t-1 + N2Yj,t-1 + N3Xi,t-1 + Ii+ t + E (7)
where Growtht is the growth rate attained from year t through year t+1. All
other
variables are as defined in equation (2), frequency weights are included and
bootstrapping techniques are once again utilized. If foreign portfolio
investment is
beneficial to firm growth, then the coefficient of FPI, or N1, should be
positive, reflecting
in increase in the growth rate with an increase of the cumulative foreign
portfolio capital
flow of the previous three years.
19
III. Data
I obtain my data for this work from the SDC Global New Issues database for
the
time period 1/1/1996 through 3/31/200311. Global new issues for all
countries are not
readily available proceeding this era in SDC. Following Korajczyk and Levy
(2003), I
exclude financial services due to the special circumstances of their asset
base and utility
firms (Macro Industry: Financial Services, Real Estate and Energy and Power)
due to the
abnormal stability and predictability of cash flow. I also exclude those firms
that have
gone bankrupt due to the special set of issues that are included in capital
structure
determination when a company is failing12. This follows the methodology of
Asquith,
Gertner and Scharfstein (1994) who found that such situations generally
cause a major
restructuring of capital structure outside of the scope of financial constraint
relaxation.
Lastly I exclude IPOs. Welch (2004) finds that the firms who undertake IPOs
find
themselves in unique environment, which contains a different set of issues
than the post-
IPO period. Including these firms would bias the results.
I collect observations for common stock, non-convertible debt, convertible
debt,
non-convertible preferred stock and convertible preferred stock issued
domestically only.
The exclusion of international issuances is intentional due to endogeneity
between
foreign portfolio investment and international issues. Financials for the
companies
issuing domestically are hand-collected from REUTERS. This approach
enables me to
have a much richer sample of global new issues around the world of firms
than afforded
me by SDC Platinum alone. REUTERS provides financial information on all
publicly
traded firms for the majority of countries in the world and as such does not
suffer from
11 Data before the beginning date of this period is sporadic.
12 Firms going bankrupt would have additional difficulty obtaining capital,
which would confound results.
20
the bias toward large firms to the extent that other international databases
such as
Worldscope/Datastream/Research Insight do. In fact, REUTERS even covers
pink sheets
and OTC/Bulletin Board firms whereas the others do not. As such, the
coverage is much
more comprehensive (see figure 3). The only firms not covered in REUTERS
are those
that have gone bankrupt or have merged with another firm. The first group
has
deliberately been excluded from the sample as previously mentioned above.
The second
group would only be a problem if the issuing company had acquired a firm in
the sense
that the capital structure control variables in these observations will have
different
relationships with the dependant variable than the remainder of the sample.
Due to the
omission of these groups of firms, there exists some survivorship bias in my
sample.
Worldscope, Datastream and Research Insight are used to confirm
accounting
values and to append the sample where available. The 31,929 observations
represent
issues of equity, debt (either convertible or straight), or preferred equity
(either
convertible or straight) and the relevant financial environment around which
the company
makes its decision regarding type of security to issue. Including the time
series of these
capital issuances brings my sample to over 106,000.
I further collect data on firms not issuing capital during this period of time to
represent those public companies that either cannot issue capital or have
sufficient funds
internally. For less developed country firm-year observations, I collect the
financials for
1996-2003 for the most exhaustive list of firms for each country as possible
from
REUTERS, collecting the exact same data utilized for the issuer dataset.
Developed
country firm-year observations are collected from Worldscope, due to the
inability of
REUTERS to provide such large amounts of data given the fact that it is
intended for
21
practitioners researching only a few companies at a time. I believe this does
not cause a
bias due to the careful matching of accounting information. Including these
non-issuer
firm-year observations, the number of observations in my dataset totals
approximately
195,000 firm-year observations.
Seven countries out of the original 53 were dropped due to insufficient
data13. In
these cases there were only one or two observations of capital issuance, not
enough from
which to obtain any statistically significant results. Two more countries
(Taiwan and
Bermuda) fall out due to insufficient macroeconomic data, leaving the
sample number of
countries to be 44. The exclusion of these countries decreases the sample
size by 3294
firm-year observations, which is less than two percent of the overall sample.
Given the fact that there are over 24,000 firms in my sample, it is not
surprising
that the range of firm-level statistics such cash, leverage, profitability and
risk is
considerable. Not surprisingly, small firms seem to have much more leverage
than their
larger peers (CDLR (2004); Rajan and Zingales (1995)). Profitability and risk
for the
small firms are considerably larger, reflecting the higher growth rate of the
small firms
(and based on the fact that the figure is scaled by assets, controlling for
size). Market
capitalization ranges from 97 (Bolivia – U.S.$MM) to 16600 (U.S. - $MM).
Annual net
foreign portfolio investment scaled by gross domestic product ranges from –
157M
(Germany) to $437B (U.S.). A full list of summary statistics for the dataset is
provided
in Table IA.
[Insert Table IA here]
13 These countries are Costa Rica, Czech Republic, Iceland, Luxembourg,
Papau New Guinea, South Africa
and Bangladesh.
22
Descriptions, as well as sources, of both firm-specific and macroeconomic
variables as well as definitions of financial data used in the analysis are
provided in
Appendix A. Size, country development and geographic distributions for the
entire
sample, as well as correlations for the variables used in my analysis are
provided in
Tables IB - IF.
[Insert Tables IB-IF here]
A. Firm-specific information
Databases, such as REUTERS, obtain financials for these listed companies
from
the exchanges. To the extent that these different exchanges in the different
countries
have different reporting requirements, financial definitions may vary.
Differences in
currency value are avoided by using ratios, which will be comparable across
countries.
This is executed through a scaling by total assets unless otherwise noted.
As many empiricists have attributed size as a determinant of capital
structure, I
assign size categories based on Total Assets. Korajczyk and Levy (2003)
and Baker and
Wurgler (2002) find a positive relationship between leverage and size.
Titman and
Wessels (1988) find that size influences not only the extent of leverage but
also the type.
My proxy for this follows both Titman and Wessels (1988) and Rajan and
Zingales
(1995) and is calculated as total assets14.
Profitability of firms would be an obvious influence on firms inasmuch as
this
impacts how well a firm could either pay interest and/or dividends. Titman
and Wessels
(1988) provide two measurements for this variable that are fairly applicable
universally.
They are operating income divided by sales and operating income divided by
total assets.
23
I utilize both in my analysis but provide results for profitability based on sales
only for
the sake of brevity.
Also relevant to capital structure determination is Asset tangibility. This
refers
to how palpable the assets of a firm are and relates to capital structure
concerns through
its limitations on debt levels due to the ability to provide collateral. A firm
has less
collateral the less tangible its assets are. This, arguably, could be said to
increase the
probability of bankruptcy due to the inability to obtain funds when there are
especially
needed. This follows logically from the fact that a company without material
assets
would not be able to liquidate to obtain the necessary funds to pay off
debtors if it were
necessary. This variable is created by calculating fixed assets divide by book
value of
assets (following Rajan and Zingales (1995)). Once again, within-country
industry
averages are used in those cases where there is missing data. For the same
reasons given
above justifying the rationale for industry average substitution as proxies for
uniqueness
of assets, industry averages are suitable proxies here.
Similar to profitability, the Growth of a firm impacts how well a firm is able
to
pay interest and/or dividends and is a typical capital structure determinant.
Proven to be
an influential variable in capital structure (Jensen and Meckling (1976);
Titman and
Wessels (1988); Chaplinsky and Niehaus (1993)), I include a proxy as
calculated by the
percentage change in total assets and also in sales revenue for a one-year
term.
To correct for any additional access a firm might have in other nations which
might affect financial constraints (Lins, Strickland and Zenner (1999)) it is
vital to
include an indication of whether a firm has listings in other countries (i.e.
ADR on a U.S.
14 This is done annually so that firms may switch size groupings over years.
The analyses are also done
using average size of the eight year periods. As results are unchanged, they
are omitted for brevity.
24
stock exchange). Including a dummy variable for Crosslisting that takes on
a value of 1
if a firm is listed on an exchange outside of its nation of domicile and 0
otherwise. I run
equation 3 for a second time, this time including the crosslisting dummy to
ascertain the
impact in light of any additional sources of capital.
B. Industry information
Differences in industry classification are avoided by using as industry
indicator
the SDC Platinum Macro industry code as my categorization. An industry
dummy is
included to account for any industry fixed effects.
C. Macroeconomic information
Based on results from such papers as Booth, Aivazian, Demirgüç-Kunt and
Maksimovic (1999), Welch (2004), and Nejadmalayeri (2001) I include
macroeconomic
factors to capture their impact on capital structures in different countries. All
macroeconomic variables, unless otherwise stated are averaged over a
lagged three year
period to abstract from business cycle effects.
GDP growth is the percentage growth in gross domestic product per capita
is
included to control for the size and development of the country.
To control for the impact of other potential sources of funds for firms I
include
both savings and foreign direct investment. Savings is the calculated as the
difference
between gross domestic product and consumption. Foreign Direct Inv., or
foreign direct
investment, is included to control for the effect provided by the more stable
of the two
global capital flows on capital issuance. This is important given the fact that
the impact
of foreign direct investment is likewise beneficial for alleviating financing
constraints
(HML (2004)).
25
Controlling for the investment environment, I include Invest, Law and
Corruption, which are indices reflecting the investment environment
attractiveness, the
level of legal development and the level of corruption (respectively) in a
country by the
International Country Risk Guide. Including proxies for the extent to which a
country’s
investment environment attracts investors, that laws are developed and that
the level of
corruption existing in a country follows the methodology of BDM (2002) and
BDL
(2003), as well as many other examinations of access to finance in an
international
setting. Papers such as Claessens and Laeven (2003) and LLSV (1997) point
out the
importance of investment climate as a determinant of financial development.
The variable of interest in this study, foreign portfolio investment, is included
in
its net form (inflows minus outflows) for the countries in the sample. Actual
levels of
foreign portfolio investment scaled by the country’s GDP and are reflected in
the variable
FPI. These scaled values are use to illicit predicted values of scaled net
foreign portfolio
capital flows based on the work of Agarwal (1997).
Instruments of the variable of interest are included due to the endogenous
nature
of foreign portfolio investment. The variable Relative Interest Rates is
included given
the potential demand for foreign investment in certain countries based on
the return
available for investment relative to other countries providing implications on
both
domestic economics and international business (Samak and Helmy (2000)).
Share is
included to address both timings of issuance that may occur (this is generally
in more
developed capital markets) or the decrease in the cost of equity and the
resulting increase
in the price of existing shares that theoretically occurs when market
integration happens
(Patro and Wald (2004); Henry (2000)). Providing an additional determinant
of
26
international trade/investment, I include 3FXRate, to provide a meaningful
value
indicator of capital investments (Agarwal (1997)). The variable of interest in
this study,
foreign portfolio investment volatility, or FPIVol, is included in log difference
terms and
scaled by foreign portfolio investment levels for the countries in the sample.
These
scaled values are use to illicit predicted values of scaled net foreign portfolio
capital
flows based on the work of Agarwal (1997). Bekaert, Harvey and Lundblad
(2003)
(henceforth these authors are referred to as BHL), Henry (2000), Patro and
Wald (2004)
and a host of other papers addressing the multi-faceted environment, as well
as the
impacts, of Liberalization. To control for these potentially confounding
impacts of
liberalization I include a dummy variable which takes on a 1 if it is included in
the
official liberalization dates of Bekaert, Harvey and Lumsdaine (2002) and/or
Henry
(2000) and a 0 otherwise.
Analogous to Kashyap and Stein (2000), I use proxies for the level of bank
lending to test the impact of an increase/decrease in the ‘money supply’ (in
my case
foreign investment). Specifically, Domestic Credit from banks and
Domestic Credit
are utilized. To test the other theory discussed in the aforementioned paper,
the ‘balance
sheet channel,’ I use the Liquid Reserves to Assets Ratio. These
variables serve as
acceptable proxies for the theories mentioned and are used instead of firm-
level data due
to the lack of loan-level data available and provide a macro-level proxy of
the same.
Fiscal Burden, from Heritage Foundation is used to control for the tax
implications of
debt in these indirect route of foreign portfolio investment analyses. This
variable takes
into consideration the proven relationship between taxes and lending in a
multinational
setting (Desai, Foley and Hines (2004)).
27
The basis of nation type for the analysis, Property Rights, is utilized to
examine
the ultimate impact of the importance of the development of the same. It
would seem
reasonable that without developed rights and the availability of recourse,
investors would
not consider investing in risky firms, small firms ranking among the top of
them. This
variable is chosen based on literature finding the importance of security law
and investor
protection such as LLSV (1997),(1998) and La Porta Lopez-de-Silanes and
Shleifer
(2004).
D. Data Correlation
Table IF provides a correlation matrix for all of the variables used in the
analysis.
There are no notable significant relationships in the firm-specific data. The
only
variables that exhibit any correlation are some of the macro variables. The
correlation of
several macroeconomic variables is significant, which is generally an issue in
many
international studies. As a result, empirical examinations using different
specifications
including select macroeconomic variables and the subsequent addition of
problematic
variables are executed to provide robustness to the results given the
potential empirical
biases based on correlation between the macroeconomic independent
variables.
[Insert Table IF here]
IV. Results
A. The Capital Issuance Choice
The results of the analysis overall support the contention that foreign
portfolio
investment assists firms in easing their financial constraints. Looking first to
LDPR
nations in Table II, we see that there exists a positive impact of foreign
portfolio
investment. Small firms in LDPR nations see a statistically significant
increase of on
28
average 2.76% in the probability of issuing capital for a one percent increase
in foreign
portfolio investment.
[Insert Table II here]
In DPR nations, the impact of FPI is also positive and significant. Small firms
in
nations with more developed property rights see on average a 0.922%
increase for a one
percent increase in FPI. This figure is not only statistically significant, but also
economically large – especially when one considers the fact that increases in
foreign
portfolio investment in the term I examine have been as much as 17%
(Ireland in 2000)
for these DPR nations. It is worth mentioning that my results in these nations
for foreign
direct investment are counter to those found in HML (2004)15 where the
sample
examined includes larger firms. The results using only small publicly-held
firms finds
that the relationship between foreign direct investment and capital issuance
in DPR
nations is instead one of a financial constraint increase. The marginal effect
on foreign
direct investment for firms in these nations seems to be significantly
negative, implying
that form of foreign investment either crowds out small firms by entering
into this type of
investment only with large firms (De Backer and Sleuwaegen (2004)), or that
any small
firms that do enter into these types of financial arrangements do so as a
substitute of
capital issuance in the public markets. Appealing to our intuition is the fact
that domestic
credit exhibits a negative association with capital issuance, demonstrating
that bank credit
is a substitution for accessing capital from the public markets. Savings also
exhibits a
positive significant association with capital issuance demonstrating that this
is yet another
source of capital which provides liquidity for a nation.
15 Performing the relevant analysis on the larger firms in my sample results
in similar findings to that of
HML (2004)
29
Importantly, I find that investment environment variables are influential in
capital
issuance. Not surprisingly, supporting the findings of such papers as LLSV
(1998), the
development of law and order is positively and significantly related to access
to finance
in DPR nations. Supporting LLSV (2000) is the positive and significant
coefficient on
corruption across property rights development. The marginal impact is
greater in LDPR
nations implying that a larger benefit may be derived in improving access to
finance in
these nations with a corresponding decrease in corruption.
Overall, firm-level variables, exhibit the expected marginal coefficients.
Intuitively appealing is the fact that in DPR nations, variables such as
leverage and cash
are significantly positively related and risk and profitability are significantly
negatively
associated with capital issuance. In LDPR nations there are fewer significant
firm-level
variables, suggesting that access to capital has less to do with these
characteristics and
more to do with country-level influences as well as infrastructure. Leverage
and cash are
exceptions to this, exhibiting a significant positive and negative relationship
with capital
issuance respectively.
These impacts, taken collectively imply that foreign portfolio investment
reaches
small firms in both DPR and LDPR nations directly through the capital
market,
effectively easing the financing constraints of firms who face difficulties
accessing
finance16.
B. The Debt/Equity Choice
Examining the second step in the capital issuance process reveals many of
the
previously unearthed contentions in empirical corporate finance literature.
Indeed, I find
16 These results are robust to clustering around industries. These results are
left out for the sake of brevity.
30
that in general, variables such as cash and leverage increase the probability
of equity
issuance, the former likely due to its evidence of liquidity and the latter
supporting the
contentions of capital structure theory. Variables such as risk and asset
tangibility
decrease the probability of equity issuance the former supported by theory
based on
information asymmetry (see, for example, Barclay and Smith (1995)) and the
latter since
this proxies a sort of collateral, against which debt may be waged. As these
variables
serve to provide control for extant literature and exhibit expected
relationships with the
dependent variable, they are left out of the tables for brevity.
With regard to the variable of interest, Table III demonstrates that foreign
portfolio investment, in general, does not seem to assist small firms to
increase their
probability of issuing equity. In fact, the probability of small firms in DPR
nations is
virtually unchanged with a 1% increase in foreign portfolio investment. This
is not all
that surprising when considering the inability of small firms to access equity
capital in
general. Foreign investors would be no more likely to extend equity capital to
informationally opaque firms than domestic investors, especially since the
vast majority
of the additional investment is in the form of debt (Henderson, Jegadeesh
and Weisbach
(2004)).
[Insert Table III here]
Keeping in mind the significant impact of foreign portfolio investment on
capital
issuance for small firms in both DPR and LDPR nations, these results reflect
the actual
breakdown of the portfolio investments, i.e. debt versus equity. Although
foreign
portfolio investment assists small firms in accessing finance, equity capital
seems to be
confined to larger firms. This further supports the contention that small firms
‘take what
31
they can get when they can get it’ since these debt-laden firms would most
benefit from
obtaining equity. It is worth noting that evidence of an international
extension of
Korajczyk and Levy (2003) is found in the coefficient on GDP growth in DPR
nations.
This negative and significant relationship with equity issuance supports their
findings -
specifically that constrained firms issue leverage procyclically.
C. The Impact on Credit Availability
Statistically significant evidence consistent with an enhanced liquidity of
banks is
found in DPR nations (see Table IV). The amount of domestic credit however,
both in
terms of banks and all financial institutions, is relatively unaffected by FPI in
these
nations. Collectively, this impact demonstrates mixed results about
extending the bank
lending channel of monetary transmission (Kashyap and Stein (2000)) to
money shocks
generically. The increase in the liquidity of bank balance sheets could imply
that banks
may be able to ‘risk shift’ their portfolio to include more risky holdings,
potentially
including more loans to small firms. This effect combined with the results of
Tables II
and III wherein small firms in DPR nations are more able to access capital in
general but
less able to access equity capital given they are able to issue, suggests that
these firms are
able to obtain public debt from the capital markets and perhaps additional
sources of bank
credit since the banks that are more able to lend credit (given the very
modest
enhancement in credit levels) are the very banks lending to small banks
(Kashyap and
Stein (2000)). The impact in LDPR nations is almost a mirror image to that in
DPR
nations - credit levels increase insignificantly while liquidity of bank balance
sheets seem
to decrease, also insignificantly. These results are not surprising once we
consider the
volatility of these capital flows and the less supportive property rights and
investor
32
protection in these nations. It appears that the same investment
environment that limits
the amount of foreign portfolio investment that enters a nation forbids any
benefits that
foreign portfolio investment capital flows might offer through the banking
channel.
Additionally, banks in these nations are often very closed to outside
investors and
corruption within this channel is omnipresent.
[Insert Table IV here]
Results at the firm level suggest that short-term and total leverage
significantly
decrease with an increase in FPI in DPR nations. This supports the
substitution effect
and lends support to the notion that with enhanced access to finance, small
firms would
not be as dependent on bank lending. Results in LDPR nations put forward a
decrease in
short-term debt but it is statistically insignificant, suggesting that these firms
remain more
dependent on bank credit. The long-term debt, however, exhibits a
significantly positive
marginal effect whose magnitude is significant as well. This likely
complements the
results in Tables II and III, which demonstrate that increases in FPI enhance
access to
finance through increased access to public debt instruments.
Collectively these results imply that foreign portfolio investment mainly flows
through the capital markets. Any impact through the indirect route of bank
lending is
very modest so that the results are seen mainly insignificantly.
D. Growth
Table V demonstrates the results of the analysis of foreign portfolio
investment
and firm growth. Not surprisingly, factors such as cash stock, profitability and
GDP
growth play an important positive role in firm growth. Leverage seems to
enhance firm
33
growth in more developed nations only implying that debt in firm capital
structures can
be utilized as a tool for growth only in relatively stable environments.
[Insert Table V here]
The growth of small firms is found to be significantly positively associated
with
FPI in DPR nations only, as demonstrated by a statistically significant
marginal
coefficient of 1.297% for these firms. It is important to remember that the
cumulative
effect given a substantial increase in FPI that can be achieved in these
nations implies
that foreign portfolio investment may provide a very real benefit for small
firms in DPR
nations with regard to growth. A positive and significant coefficient on
leverage in DPR
nations (not shown) offers a nice explanation as to how this growth is
achieved when
coupled with the increased probability that small firms will issue debt should
they issue
capital. Small firms in LDPR nations also see a significant increase in growth
of total
assets with a 1% increase in foreign portfolio investment of on average
2.56%.
Growth enhancement in terms of sales revenue is only apparent for firms in
DPR
nations. Small firms in developed property rights nations see an average
1.501% increase
with a 1% increase in sales revenue growth. Results for LDPR nations with
regard to
sales revenue growth compliment an IMF study that finds that foreign
portfolio
investment has not benefited growth at the country level in developing
countries. Some
might argue that sales revenue is the more meaningful measure of growth
since growth in
total assets in not necessarily value creation (e.g. in the case of “empire
building”).
The significantly enhanced growth rates in terms of both total assets and
sales
revenue for firms in DPR nations are important in light of the controversy
around foreign
portfolio investment and its short-term nature. Much of the literature
demonizing this
34
tool of globalization has implied that the volatility implicit in this form of
investment
must ultimately be harmful for productivity. At the firm level, this analysis
casts doubt
on this contention given a sufficient level of property rights is present. This
provides
hope for small firms and should motivate top officials to increase the level of
property
rights in their domicile nations and to further encourage firm-level
improvements in
corporate governance/disclosure.
V. Robustness
A. Financial Constraint Definition
To provide an additional test of the impact of foreign portfolio investment on
the
financial constraints of all constrained firms, I need to first determine the
external
financing each firm needs. BDM (2002) finds that firms seem to employ more
long-term
capital for growth in market-based environments. It is from this paper that I
take my
methodology for discerning financial need, or external funds necessary. They
derive
their dependent variable from the “percentage of sales” approach to
financial planning17,
calculating the external funds necessary. I use a slightly different version of
the same
equation to estimate my need for external financing. Based on these results,
I further
include proxies for both market and bank development. The altered ‘external
funds
necessary’ (or EFN) is calculated as follows:
EFNt = (At /St)(St – St-1) – (Lt /St) (St – St-1) – Mt (St)(RR) (8)
where At is the total assets of the firm in time t, St-1 and St are the sales of
the firm in
times t-1 and t respectively, Lt is the liabilities of the firm in time t, Mt is the
profit
17 Demirgüç-Kunt and Maksimovic (1998) also uses this methodology.
35
margin of the firm as defined by net income divided by sales for time t, RR is
the
retention ratio for the firm. As noted by BDM (2002) several simplifying
assumptions
are made in order for this methodology to be implemented. First, both the
asset
utilization (A/S) and the profit margin of the firm must remain constant per
unit of sale.
Further, the use of the formula to discern additional funds necessary
depends on true
values of assets being reported (relative to their depreciable basis).
I adopt the Rajan and Zingales approach (Rajan and Zingales (1998)) to
obtain
unconstrained growth rates for the sample since using firm-specific
information would
imply that the resulting predicted growth rates would be optimal18. The
growth rates of
the relatively unconstrained firms in the US are mapped by both industry and
size to all
other countries and these observations are subsequently dropped from the
analysis19.
Beyond industry matching, firms are matched by size due to the fact that
small firms in
countries with even the most developed property rights have difficulty
accessing capital
(BDM (2005); Koraczjyk and Levy (2003)). This methodology avoids any
downward
bias based on access to capital and any bias for analysis on US firms after
using these
firms to obtain the unconstrained growth rates.
Dropping those firm-year observations where EFN is negative and testing the
primary hypothesis of the paper once again (H1) yields very similar results20
with regard
to sign. Magnitude of the impact seems to be enhanced considerably
however for firms
in DPR nations. This would suggest that there were firms in the previous
sample who
perhaps did not need financing, diluting the effect such that magnitude was
decreased.
18 Rajan and Zingales examine inter-country differences between industries
based on both macro and firmspecific
information
19 This results in 52, 276 observations being dropped.
36
Small firms in LDPR nations unfortunately do not see the same
enhancement, suggesting
that relative to their larger counterparts, these firms do not benefit as much.
Table VI
Panel A shows these results.
[Insert Table VI here]
B. Measurement of FPI Flows
Inasmuch as the portfolio investment cash flows given in this analysis are net
(i.e.
cash inflows minus cash outflows), an increase in foreign portfolio
investment does not
necessarily imply an increase in domestic investment. To infer when foreign
portfolio
investment inflows increase, I test sub periods of the data. IMF’s Coordinated
Portfolio
Investment Survey (2001) reports the previously mentioned trend toward
investment in
emerging markets during the 90s and later the mass exodus of foreign
investment capital
flows changing the percent invested in developing nations from 65% to 9%
(see figure 4
and 5). Foreign portfolio investment continues to grow but thus far this
decade, it
appears this increase in investment occurs in developed nations. As such, I
provide tests
on time subsets of data for different market capitalization groups. For less
developed
nations, I test equation (2) and (3) using the years 1996-1999. For more
developed firms
I test the same equations over the term 2000-2003. In so doing, the results
remain. Not
surprisingly, they are even enhanced by this investment trend horizon for
LDPR nations.
This could imply that in these good times for hot money, small firms in these
nations
have a fighting chance to access capital. Results may be seen in Table VI
Panel B.
20 Only approximately 25% firms dropped using this methodology issued at
some point during this eightyear
term, compared with approximately 44% of firms with a positive EFN.
37
C. Interactions with Investment Environment Variables
Examining the effect interactive effect of FPI with investment environment
variables gives us a better idea of the cumulative effect of FPI on small firm
access to
capital. Examining the implications of FPI interacted with these variables on
the entire
dataset (across firm size) offers further insight that investment environment
is important.
This should be fairly intuitive given the results of the vast literature by such
authors as
LLSV, Bekaert and Harvey and Wurgler. The impact of FPI including
investment, a
variable that refers to a governments attitude toward foreign investment, is
one that
implies decreasing margin to return. For those governments at the bottom of
the index
(invest=0), the marginal impact of FPI on access to capital is large – 1% for
each percent
increase in FPI. For those governments already open to FPI, the marginal
effect of
becoming more open is less. Those governments with that investment index
equal to
twelve, for example, would see only a 0.766% marginal benefit of FPI on
access to
capital. The impact of FPI given law and order is also important. Although its
magnitude is less than that of investment, it possesses increasing returns to
scale. Firms
in nations where this index is low (law=1) can expect a 0.327% increase in
access to
capital with a 1% increase in FPI. Those nations where law and order is more
developed
(e.g. law=6) will see a more impressive 0.707% increase. Lastly, the impact
of FPI given
corruption seems to be quite influential. The cumulative impact suggests
that the impact
of FPI in the presence of this variable is pivotal. For example, looking to
those nations
where corruption is rampant, the cumulative impact of FPI is decidedly
negative for firm
access to capital - (-)4.736%. In those nations where corruption is well under
control
(e.g. corruption=6), the impact is equally large with regard to magnitude but
this time
38
positive – (+)4.394%. These results suggest that although all three
investment
environment variables are important, corruption levels within a nation can
determine
whether the marginal impact of FPI is positive or not. Results may be found
in Table
VII.
[Insert Table VII here]
D. Alternate Definitions
Performing sensitivity analysis around definitions of key variables such as FPI
and firm size, as well as altering sample country inclusion definition provides
some
robustness for the results. To use another definition for foreign portfolio
investment, I
scale the net flow by gross private capital flows into a nation instead of the
previous scale
– gross domestic product. This is the FPI variable utilized in HML (2004).
Instead of
defining size annually, I instead use average size over the term examined. To
alter
sample country inclusion specifications, I drop countries that may bias
results due to
changes in capital control policy or specific laws which may bias results such
as in China,
where only B shares were offered on the market for foreign investors during
this term and
foreign banking was not possible before 2002. These countries include China,
Malaysia,
Hong Kong, Korea and Chile21. Performing these three specifications
changes things
only slightly.
The utilization of the definition of FPI from HML (2004), scaling by total
private
global flows instead of GDP, proffers the main difference from previous
results. This
point is brought out earlier in the methodology section of the paper and
supports the
21 Malaysia had capital controls until 10/1/1998 and South Korea was
liberalized in 1998, which is two
years after the first year of the examination period. Chile initiated the
encaje, which is legislation that may
have had an impact on FPI levels and Hong Kong did not have FPI levels for
part of the sample period.
39
reconciliation of the differences in results in this work and those of HML
(2004).
Overall, results remain similar and can be found in Table VIII.
[Insert Table VIII here]
VI. Chapter Conclusions
Small firms play a distinctive and influential role in both the present and the
future economic situations in which nations find themselves. Financial
constraints for
these firms are exacerbated by both firm- and macro-level influences, and as
such,
additional sources of investment as potential additions to accessible finance
are worthy of
investigation. Examining the importance of foreign portfolio investment in
the capital
issuance process, I find that foreign portfolio investment enhances the
accessibility of
investment capital through financial markets for small firms in countries
across property
rights development, but through the bank lending channel only in those
countries where
property rights are more developed. Further, the enhanced access to capital
for small
firms only leads to significant value-enhancing growth in those nations that
have more
developed property rights. The fact that small firms in nations with more
developed
property rights can look forward to enhanced firm-level growth with
increases in foreign
portfolio investment underscores the importance of investment environment
fundamentals. Improvements in a country’s foreign investment environment
serve to
increase the probability of financial constraint alleviation, for small firms in
less
developed nations.
The positive influence of foreign portfolio investment on small firm access to
capital supports the ideals of those who strive for optimal policy reformation
in those
nations who do not support foreign investment and whose markets are
excessively
40
volatile or underdeveloped with regard to investor property rights. Easing
foreign
portfolio investment restrictions on capital flows; stabilizing these
investment cash
inflows and improving the treatment of foreign companies and investors
could have a
very real influence on the longevity of the small firm.
41
Chapter 2: Taking the Bad with the Good: Volatility of Foreign
Portfolio Investment and Financial Constraints of Small Firms
I. Motivation
A. The Benefits of Market Integration
Research done at the macro level shows that liberalization of investment
regulations reduces the cost of capital in a country through capital market
integration,
increases capital flows such as foreign portfolio investment into the host
country (Bekaert
and Harvey (2003)), increases stock returns during the process (Patro and
Wald (2004))
increases the liquidity and size of markets (Levine and Zervos (1996)), and
leads to an
increase in the real economic growth over a medium-term (BHL (2003)).
Focusing on
the stock market impacts mentioned, the supply-side of capital increases,
and the
increased depth of financial markets caused by the level of foreign portfolio
investment
flowing into a financial market potentially eases the financial constraint of
firms (Laeven
(2003); Chapter 1)22, improves the allocation of capital (Wurgler (2000)) and
importantly
is often accompanied by improvements in transparency of accounting
reporting and
corporate governance (Feldman and Kumar (1995)).
Importantly, the desire of countries, and the companies within them23, to
“pull”24
foreign portfolio investment to their economies motivates improvements in
such things as
22 See also Chari and Henry (2004) who find that the growth rate of a firm’s
capital stock exceeds that of
its pre-liberalization rate.
23 A McKinsey & Company Global Investor Opinion Survey (2002) finds that
investors are prepared to pay
a premium for companies exhibiting enhanced corporate governance
standards. This premium is on
42
corporate governance (Shinn (2000)) and investor protection/property rights
(Bekaert and
Harvey (2003)). This in turn leads to increased investment (Dahlquist,
Pinkowitz, Stulz
and Williamson (2002); Claessens and Laeven (2003))25 and a cycle of
investment
environment improvement ensues. This cycle is longer-term in nature and is
not likely to
stop suddenly based on changes in the level of foreign investment.
B. The Costs of Market Integration
The impacts of the influx of capital flows and capital market integration that
occur in such events as liberalization, however, are not necessarily all good.
Bekaert and
Harvey point out in their liberalization literature survey that liberalization,
and the
resulting increase in investment capital, may also have negative impacts.
These negative
impacts are vastly due to the short-term nature of this capital flow - its
volatility, and the
potential to cause investor unease or panic. This panic can either be a result
of, or caused
by, crisis within a country. In contrast to the capital flow level, the variance
of the flow,
sometimes referred to as its instability, causes pressures on the money
supply, exchange
rates and stock market volatility26 27 of its host nations, making keeping
tight reigns on
economic policy difficult for governments and at times arguably increasing a
country’s
propensity for crisis or exacerbating the impact of an existing crisis. Henry
(2003) points
out that crises such as those in Asia, Russia and Latin America have
challenged the merit
of capital-account liberalization. Henry (2000) questions the permanency of
the increase
average 12-14% in North America, and Western Europe, 20-25% in Asia and
Latin American and over
30% in Eastern Europe and Africa.
24 See, for example, Calvo, Leiderman and Reinhart (1993); Claessens
(1995); Claessens, Dooley and
Warner (1995)
25 The opposite effect is also true – see Aggarwall, Klapper and Wysocki
(2003) for characteristics
investors look for in foreign investment countries.
26 See also Patro and Wald (2004).
27 Bekaert, Harvey and Lundblad (2004) however contends that this actually
that the data do not support
this.
43
in capital, suggesting that the increase in liquidity may only be temporary.
Henry and
Lorentzen (2003) differentiate between liberalization with regard to equity
and debt,
stating the latter can be dangerous since it leads to a reliance on debt in the
capital
structure. This is provocative given the fact that debt comprises about 90%
of new
capital issued internationally around the world (Henderson, Jegadeesh and
Weisbach
(2004)). Demirgüç-Kunt and Detragiache (1998) find that banking crises are
more likely
to occur in liberalized economies. This is relevant to small firm access to
finance not
only due to the frequency of twin crises, but also because crises in the
banking sector
could devastate small firms due to the fact that bank loans are the mainstay
of their
financing.
Given that small firms are so very sensitive to macroeconomic conditions
(BDM
(2002); Tewari and Goebel (2002)), increased volatility could diminish any
benefit
achieved through the increased supply of investment capital. Indeed, Samak
and Helmy
(2000) find in their examination of foreign portfolio equity investment in
Egypt that
maximizing the ultimate value of this form of foreign investment is
dependent upon
macroeconomic stability and a strong existing market infrastructure. To
complicate
matters further, the areas that seem to have the most to gain from global
investment
capital flows such as foreign portfolio investment seem to enjoy these capital
flows only
accompanied by potentially damaging capital flow volatility (see figure 4 and
5).
C. Weighing the Impacts of the Capital Flows and its Volatility
Whether or not the potentially damaging aspect of FPI, volatility, overpowers
the
benefits derived from the actual capital flow itself (e.g. increased liquidity,
improved
allocation of capital, improved corporate governance/investor
protection/transparency)
44
which could ultimately reverse the enhanced small firm access to capital,
depends upon
the impact of FPI volatility on these benefits. Investment environment
improvements
such as corporate governance, investor protection and/or transparency are
put into effect
because of capital inflow volatility and are supported by the company trying
to obtain
financing, government officials trying to attract foreign investment, foreign
investors
with a potential stake in their investment28 and official aid organizations
such as the
World Bank with the intent to decrease the volatility of capital flows.
Examples of
government legislation requiring these improvements in
disclosure/transparency as well
as improved corporate governance in less developed nations are
increasing29. These
laws seeking to improve corporate governance and indirectly investor
protection also
seek to stabilize capital inflows, making it less likely to be positively
correlated with FPI
volatility or crisis. Improvements of corporate governance at the firm level,
induced by
competitive forces for capital – both domestic and foreign - are not likely to
be dropped
by firms simply because their domicile nation is in crisis or that FPI becomes
more
volatile perhaps even leaving the country for a year or two. In fact, this
might induce
firms to improve corporate governance measures such as board of director
composition or
disclosure even further, or at least to maintain the improvements already
made to attract
future foreign capital and to maintain or establish better access to capital
domestically.
The benefits of FPI may actually serve to ultimately decrease a country’s
dependence on
foreign investment by improving the investment environment enough to
stabilize
domestic investment which will eventually decrease the damaging impacts
of the
28 See Khanna and Palepu (1999).
29 Korea has implemented a law requiring domestic companies to produce
quarterly results. China is
switching from “cash” to “accrual” accounting. Brazil has just legally limited
the number of non-voting
45
volatility of these capital flows. Assuming this is true, any potentially
damaging effects
of FPI could be attributed to “short run pain for long run gain30.”
Although liquidity is more short-term in nature and would likely be impacted
by
volatility in the level of foreign portfolio investment, the positive impact of
FPI would
only be nullified if market liquidity reverses in the presence of FPI volatility.
This does
not seem to be the case. The liquidity of markets does not systematically
decrease with
FPI volatility, as is seen in figure 6. In fact, the correlation between the two
when using
total value of listed securities traded as liquidity is a significant positive
0.505131. This
correlation as well as studies that foreign investors do not destabilize
markets any more
than local investors (Dvorak (2001)) challenges the notion that market
liquidity drops in
volatile times (e.g. the flight of foreign capital)32. If the mechanisms by
which the
majority of the benefits of FPI with regard to small firm access to capital (see
Chapter 1)
is enhanced market liquidity, and volatility is significantly positively
correlated with
market liquidity in countries considered “investment grade,” as defined as
periods when
confidence in a country’s tranquility is higher than the sample median for
that time
period, it could be posited that FPI volatility does not destroy the enhanced
access to
capital small firms achieve coincidental to FPI flows.
Given the lack of compelling evidence that FPI volatility damages or
neutralizes
the positive benefits of FPI along with the anecdotal evidence that liquidity
may not be
decreasing with volatility in all times, I contend that the volatility of FPI, as
measured by
shares a company can issue. Mexico has created a law which precludes
holding companies from gaining a
controlling share of a company to force minority shareholders to sell at below
market value.
30 See Kaminsky and Schmukler (2002).
31 Using percent of market traded instead of total value traded yields a
significant correlation of 0.3902.
32 See also Borensztein and Gelos (2001) and Karolyi (1999), who find that
the herding of investors, which
is often cited as the cause of the volatility of this capital flow, is not
significantly different in crisis versus
noncrisis periods.
46
the logarithm of the variance of FPI net flows scaled by a proxy for the size of
an
economy, gross domestic product (GDP), for the period t-1 through t-3, does
not
significantly decrease the access to finance of small firms in all times. In
periods when
foreign institutional investors have more confidence that the country is
relatively immune
to imminent crisis (i.e. lower country risk), waves of foreign portfolio
investment should
not hinder small firm access to finance. Tested empirically this becomes:
H1) Controlling for the level of foreign portfolio investment, the volatility in
foreign portfolio investment (scaled by the size of the host county) does not
significantly impede small firm access to capital, as measured by the
probability
of capital issuance, in times of increased country confidence, as measured
by an
increase in the institutional investor rating.
It is worth noting here the importance of the inclusion of the FPI level so that
the
impact of the volatility of the flow can be disentangled from the level itself.
Including
this variable should enable the true effect of the instability of this capital flow
to be
uncovered. Also relevant is the fact that volatility is scaled by gross domestic
product.
This is to address the fact that large developed countries such as the United
States
actually have a larger FPI volatility than smaller countries such as Peru, yet
they are able
to absorb such things often without negative implications.
Bekaert and Harvey, Henry (2000) and Henry and Lorentzen (2003), papers
described earlier in the motivation, point out the potentially negative
attributes of capital
flows such as increased pressure on money supply, exchange rates and
market volatility,
and mainly base these contentions on the volatility inherent in this short-
term capital
flow. Given the potentially fickle nature of this capital flow coupled with the
sensitivity
of small firms to macroeconomic volatility (BDM (2005)), would an increase
in FPI
volatility impact the growth of small firms? Even if H1 can not be disproved,
and
volatility increases the ability of these firms to raise capital in periods of
enhanced
47
country creditworthiness/low propensity for crisis, is it ever good for small
firm growth?
Given the sensitive nature of small firms to macroeconomic factors, as well
as the
negative impact of macroeconomic volatility on small firm access to capital,
it is likely
that volatility has a negative impact on the growth of these firms.
H2) Controlling for the level of foreign portfolio investment, the volatility of
foreign portfolio investment (scaled by gross domestic product) hinders the
growth of small firms as measured by the log difference in both total assets
and
sales revenue.
II. Methodology
A. Volatility in Foreign Investment
To test whether the volatility of foreign portfolio investment, as calculated as
the
logarithm of the variance of foreign portfolio investment over years t-1
through t-3 is
damaging to small firm access to capital, I divide my sample of 44 countries
into subsets
based on the creditworthiness of the country-year – “investment grade” for
those countryyears
more than the annual sample median Institutional Investor Rating and
“noninvestment grade” for those country-years less than the annual sample
median. This
is important given the fact that “shifts in international portfolio composition
usually
correspond to changes in perceptions of country solvency by international
investors
rather than to variations in underlying asset value (FitzGerald (1999)). It is
also
important given the responsibility that investors are given for their role in
crises. The
quote at the beginning of this essay from the Federal Reserve Bank in San
Francisco
points to the popular perception that investor panic causes crises - not asset
value – thus
investor perception of risk is an important factor in the potential downside of
FPI.
Inasmuch as sovereign risk is determined to be a leading indicator of crisis
(Kaminsky,
Lizondo and Reinhart (1998) – henceforth these authors will be referred to as
KLR), but
48
not a significant predictor of FPI values (Agarwal (1997)), concerns of
interdependence
between these categories and FPI volatility should be allayed. Estimating the
impact of
sustained volatility on small firm access to finance, as measured by the
probability of
capital issuance (y=1 where firm i issues capital in time t and equals zero
otherwise), I
perform the following regression.
Prob(y=1)j,t = ?0+ ?1FPIVolj,t-1 + ?2 FPIj,t-1 + ?3Xi,t-1 + ?4Yj,t-1 + Ii+ t + E
(10)
where FPIVol is the predicted level of FPI volatility from the first-stage in the
instrumental variable probit regression (see equation (11) for the first stage)
and is
calculated as the variance of foreign portfolio investment scaled by gross
domestic
product (GDP), FPI is the average level of foreign portfolio investment scaled
by GDP in
the period t-1 through t-3 (parallel to the volatility term), X is a vector of
lagged firmspecific
variables such as cash flow, debt/asset level, profitability, risk, growth,
external
financing necessary, asset tangibility and crosslisting. These variables
control for
occurrences wherein firms would be more likely to issue (see for example,
Korajczyk,
and Levy (2003), BDM (2002), Baker and Wurgler (2002)). Y is a vector of
lagged
alternate sources of capital such as foreign direct investment, foreign
portfolio
investment, domestic credit and savings. FPI, in particular, is added to
distinguish
between investment level/trend33. Macroeconomic variables are averaged
over the years
t-1 through t-3 to in order to parallel the volatility term and to abstract from
business
cycles. This methodology is often used in cross-country analyses to smooth
out annual
fluctuations that can otherwise confound results (see BDL (2003); Rousseau
and Wachtel
33 Any concerns that interdependence between foreign portfolio investment
flows and FPI volatility may
drive results should be resolved by the fact that pairwise correlation of these
two is once again below 10%
and insignificant.
49
(2002)34). I is a vector of industry dummies to control for industry effects
and t
represents time dummies, which control for any time effect in the panel. A
description of
the firm-, industry- and country-specific variables is in the data section as
well as in the
appendices35. The instrumental probit methodology used implements
frequency weights
to avoid data cloning issues and utilizes a bootstrapping methodology, which
uses
randomly chosen subsamples36 of the dataset with replacement to avoid
dependence on
assumption of the normality of distribution or the absence of stochastic
influences on the
data.
According to Agarwal (1997), the significant determinants of foreign portfolio
investment are inflation, the real exchange rate, market capitalization and
some proxy for
economic activity. Inasmuch as the actual capital flows are suffering from
potential
endogeneity issues, volatility of these capital flows will likely suffer the same.
Supporting this contention is the statistically significant correlation between
FPI volatility
and other macroeconomic variables utilized in the analysis. Since
endogeneity of the
volatility of foreign portfolio investment is a concern, I utilize an instrumental
variable
approach that in the first stage estimates FPI volatility and in a second stage
estimates the
regression in equation (10). Robust standard errors are adjusted to allow for
within firm
correlation of observations and a two-stage approach. I regress foreign
portfolio
investment instability (I define volatility in absolute, relative log difference
terms, as well
as the change in volatility to ensure robustness) on relevant variables, such
as corruption
based on the relevance found in Chapter 1, relative interest rates and
liquidity, based on
34 See also Demirgüç-Kunt, Laeven and Maksimovic (2003).
35 Tobin’s Q is not included in my analysis due to the scarcity and lack of
consistency of information on
market pricing in both less and more developed nations around the world.
36 N=50 is used for bootstrap replication.
50
work from Bekaert and Harvey (2003) and country sovereign risk as well as
foreign
exchange rate changes, based on work from KLR (1998) and Reinhart, Rogoff
and
Savastano (2003)37. The empirical model is as follows:
FPIVolj,t = F0 + F1AFXRatej,t + F2Corrj,t + F3RelIRj,t+ F4ATVTj,t +
F5AIIRj,t + t + E (11)
Performing the two-stage regression, I examine the impact of
endogenouslydetermined
FPI volatility on the probability of a firm issuing public capital. Controlling
for other influences in capital issuance, the relationship that exists between
a finite
change in volatility (increase or decrease) and the probability of capital
issuance will
offer support or help to reject the hypothesis, H1. I expect this coefficient,
W1, in equation
(10) to be positive and significant for small firms in the investment grade
sample. Small
firms in the noninvestment grade sample will likely exhibit a negative
association with
capital issuance due to the negative relationship between liquidity and
volatility in these
times and in regressions that don’t delineate between increased or
decreased country risk.
B. Growth
To examine whether the volatility of foreign portfolio investment ultimately
hinders firm growth (H2), I utilize the growth rates of these firms (in total
assets and in
sales revenue) and whether that impact is disproportionate by regressing the
following:
Growthit = U0+ U1FPIVolj,t-1 + U2FPIj,t-1 + U3Xi,t-1 + U4Yj,t-1 + Ii+ t + E
(12)
where Growthi,t is firm i’s growth rate attained from year t through year t+1.
All other
variables are as defined in equation (10). If foreign portfolio investment
volatility
hinders small firm growth, then the coefficient of FPIVol, U1, should be
negative,
37 All instruments are significant at the 1% level in this stage.
51
reflecting in decrease in the growth rate with an increase in the volatility of
foreign
portfolio investment volatility.
III. Additional Data
The collection and contents of the dataset utilized for this dissertation are
outlined
in Section III of Chapter 1 of this work. Additional data necessary for this
chapter are
listed in this Chapter/Section. Descriptions of all variables, for the cumulative
work are
available in the appendix.
The liquidity of capital markets, proxied by Total Value Traded, is included
to
instrument FPI volatility. This is due to the negative effect reduced market
liquidity has
on the confidence of foreign investors (Aggarwall, Klapper and Wysocki
(2003)).
The variable of interest in this study, foreign portfolio investment volatility,
or
FPIVol, is included in log difference terms and scaled by foreign portfolio
investment
levels for the countries in the sample. These scaled values are use to illicit
predicted
values of scaled net foreign portfolio capital flows based on the work of
Agarwal (1997).
Investment Grade and Noninvestment Grade are variables created to
depict the
environment within which investors find themselves. This classification is
created based
on a rating of the creditworthiness of the country – the Institutional
Investor Rating.
This rating is used by Reinhart, Rogoff and Savastano (2003) to infer the
general
impression of a country’s solvency with regard to foreign debt and has
implications on
how volatile short-term investment may be as a result of confidence (or the
lack thereof)
in a nation’s proximity to crisis. The relevance of institutional investor’s
impression of
the solvency of country sovereign debt has also been mentioned in Samak
and Helmy
(2000) as an important factor in the “pull” of foreign portfolio investment to
a country.
52
Lastly, I include Propensity for Crisis in the robustness section of this
chapter.
This is used as an alternative and perhaps more direct measure of a
country’s proximity
to crisis based on the works of Kaminsky (2003) and KLR (1998) which
examines the
timing and leading indicators of crises respectively.
IV. Results
A. Small Firm Access to Capital
The volatility of foreign portfolio investment could theoretically pose a threat
to
existing investors via security values and the firms via asset values if and
when capital
leaves the country very rapidly in times when investor confidence is quite
low, for
example, in crisis periods. This volatility, however, does not have to translate
into a
decreased level of access to finance if the short-term effects such as a
decrease in
liquidity does not outweigh the longer-term benefits of foreign portfolio
investment that
make this enhanced level of financing access possible, such as strides to
improve the
investment environment. In support of that logic, Table IX displays an
economically
insignificant negative coefficient for small firms in all three samples – on
average of
0.012 – that is a 0.012% decrease for a 1% increase in the level of volatility.
Even in a
multiplicative sense given probable swings in the level of volatility, this is a
very small
number. Checking the coefficient on the control for the level in FPI, it is
further apparent
that the volatility coefficient (marginal effect) is not large enough to
completely reverse
the benefits of FPI. In fact, it hinders it minimally when defining volatility in
this
manner. Firm access to capital in the noninvestment grade sample is not
even
statistically significant, suggesting that volatility in these country-years is
absorbed and
goes relatively unnoticed.
53
[Insert Table IX here]
Using a relative measure of volatility as defined as the level of volatility
relative
to global volatility provides similar results. Only the magnitude of the
marginal effects
changes, increasing to an average of 0.065, implying that it is the relative,
rather than the
absolute, level of volatility that matters with regard to the level of impact.
Having said
that, the noninvestment grade sample (specification 3) is insignificant – this
was not the
case when using a straight-forward volatility definition. This is perhaps due
to the fact
that when a country-year is deemed noninvestment grade, its volatility
relative to other
country-years is not as important as the fact that it is currently considered a
bad
investment.
Looking to the change in volatility, it becomes obvious that changes in the
variance of FPI net flows inhibits access to capital in general. Both the
economic and
statistical significance of this negative association of FPI volatility perhaps
speaks more
to proximity to periods of crises and the confidence of not only foreign
investors but also
to domestic investors. Increases in the level of variance suggest a much
more volatile
macroeconomic environment, one that would decrease the pool of
“investible” firms by
causing a flight to quality by investors to safe investments. The marginal
effect of the
change in volatility on access to public finance is negative and significant
across the
board for this definition of volatility. What’s more, the economic significance
has
increased to on average 0.127, implying a more significant effect once one
considered the
potential multiplicative effect of this coefficient given swings in FPI volatility.
Alternate sources of capital demonstrate expected relationships with capital
issuance (access to public finance). Foreign direct investment has a positive
influence on
54
capital issuance as does national savings. Domestic credit, a substitute for
public capital
issuance, is negative. Interestingly, the marginal effect of GDP growth
demonstrations
that firms issue counter-cyclically, when they are more likely to need
external financing –
in the investment grade sample. Noninvestment grade, and to a certain
extent, the whole
sample, show a positive relationship with GDP growth and access to finance
which could
imply some type of capital rationing wherein firms only receive access to
capital in better
times. Comprehensive results may be found in Table IX.
B. FPI Volatility and Firm Growth
Importantly, the results in Table X show that volatility may indeed be bad for
small firm growth but it seems as if it doesn’t always have to be. Perhaps
surprisingly,
we see that growth in the base specification (“all” times) for growth in total
assets is not
significantly negatively associated with an increase in FPI volatility and
growth in sales
revenue actually reflects a positive significant association with the same.
This is good
news for proponents of capital market integration since it implies that FPI
instability does
not have to hinder firm growth, which in turn implies that it may not derail
the economic
growth that BHL (2003) suggest might ensue with this integration.
Corresponding nicely
with the results in Table IX is the fact that growth in the investment grade
sample, growth
in sales revenue is positively impacted by FPI volatility. However, it is not
statistically
significant. The noninvestment grade sample seems to be the only
specification where
FPI volatility exhibits a significantly negative association with growth. This
impact is
not surprising given the results for these firms in Table IX demonstrating a
decreased
probability of being able to access public financing in two of the three
volatility
definitions, coupled with the typically enhanced risk aversion of investors in
these times.
55
The extent of this marginal effect relative to the benefits of FPI seems to be
more
significant than the effect on access to finance. For growth in total assets in
particular
these marginal effects, although insignificant, demonstrate a potentially
threatening force
for these small firms. Growth in sales revenue seems to offer a more
pronounced effect,
suggesting that this form of growth is more immediately reactive to cash
flow, but
reflects a much less threatening effect, offering some support to the
contention that FPI
benefits are not neutralized in the presence of its volatility. Overall, these
results should
help to allay fears that volatility hinders these firms, at least in the short run.
[Insert Table X here]
V. Robustness
A. Alternate Definitions and Sample
Performing sensitivity analysis around definitions of key variables such as
FPI, as
well as altering sample country inclusion definition provides some robustness
for the
results. To use another definition for foreign portfolio investment, I scale the
net flow by
gross private capital flows into a nation instead of the previous scale – gross
domestic
product – and calculate the logarithm of the variance of the term t-1 through
t-3 based on
this definition. This definition of FPI is utilized in HML (2004). I also define FPI
scaling by market capitalization. Lastly, to alter sample country inclusion
specifications,
I drop countries that may bias results due to changes in capital control policy
or specific
laws which may bias results such as in China, where only B shares were
offered on the
market for foreign investors during this term and foreign banking was not
possible before
56
2002. These countries include China, Malaysia, Hong Kong, Korea and
Chile38.
Performing these three specifications leaves the vast majority of the results
in place. The
magnitude of the marginal effects is slightly altered but overall, results
remain similar
and can be found in Table XI.
[Insert Table XI here]
B. Proximity to Crisis
To address concerns that the volatility measure utilized does not capture
fully the
downside of FPI, I reexamine the data using a measure which captures a
country’s
proximity to crisis perhaps more directly. Using the country-years depicted in
Kaminsky
(2003) as currency crisis years and BDL (2002) and Demirgüç-Kunt and
Detragiache
(2005) for banking crisis years, I create a dummy variable which takes on a
value of one
if a country is in either a currency or a banking crisis and zero otherwise. The
inclusion
of the banking crisis variable is due to the frequency of banking crises and
currency
crises to occur simultaneously – the so-called “twin crises” (Kaminsky and
Reinhart
(1999); Zhu (2003)). Using leading indicators of crises from Kaminsky,
Lizonda and
Reinhart (1998), I regress the following:
Prob (y=1)j,t = V0 + V1FXRatej,t-1 + V2AIIR j,t-1 + V3NetCapAcct j,t-1 +
V4Reserves j,t-1 + W (13)
where FXRate is the real exchange rate, AIIR is change in the institutional
investors’
Country Sovereign risk rating, NetCapAcct is the net capital account level,
and Reserves
is a country’s amount of reserves. I perform this cross-sectional probit
regression both in
and out-of sample.
38 South Korea was liberalized in 1998, which is two years after the first year
of the examination period.
Chile initiated the encaje, which is legislation that may have had an impact
on FPI levels and Hong Kong
57
B.1 In-sample
Looking at the in-sample regression first, the following regression is run.
Prob(y=1)j,t = N0+ N1FPIj,t-1 + N2CrisisPropj,t-1*FPIj,t-1 + N3FPIj,t-1 +
N4Xi,t-1 + N5Yj,t-1 + Ii+ t + E (14)
where CrisisProp is the propensity for a country to go into crisis as defined as
the
instrumented value, or the first stage of a two-stage least squared
regression. All other
variables are defined as previously in the paper.
Results from this analysis provide insight as to how the benefits of foreign
portfolio investment deteriorate with an increase in a country’s risk of crisis.
The
interaction term in Table XII shows that this impact is not surprisingly
negative. Taken
collectively with the positive and significant effect of the FPI term implies
that as the
propensity for crisis grows large for country j, the benefit derived from FPI
decreases. In
fact, this benefit is completely reversed when the propensity for crisis
reaches only 24%.
This fits in nicely with the volatility analysis since we see that all countries
can be
hindered by the volatility in this capital flow but not in all times. Although the
enhanced
access to finance gained from this foreign capital flow falls with an increase
in the
propensity for crisis, a positive benefit is retained for most of the sample.
Indeed, the
mean propensity for crisis in the sample is only 18.5%, indicating that this is
not the case
for the majority of the sample. The results do indicate, however, that for
those countries
particularly sensitive to crisis, enhancements in access to finance may not be
maintained
if stability in these economies is interrupted. In fact, the effect seen in the
interactive
variable coefficient relative to the crisis propensity variable alone shows us
that FPI
did not have FPI levels for a portion of the examination period.
58
actually does exacerbate the effect of the crisis – as the popular press
accuses. Although
this is not great news for advocates of market integration, it underscores the
importance
of a stable infrastructure and investment environment that will endure the
challenges that
currency and/or banking crisis offer an economy. Recognizing that the
definition of
crisis in this examination includes banking crises and acknowledging once
again that
currency crises and banking crises may well occur contemporaneously, the
banking
sector, as well as financial markets, plays a large role in the stability
maintenance of
countries39.
Results regarding growth are similar supportive of earlier findings. Looking to
growth in sales revenue, the definition of growth that many economists feel
is more
valuable to the economy, we see that although FPI is positive and
significantly related to
small firm growth, its effects diminishes when combined with a nation’s
propensity for
crisis. The sample average of 18.5% just nullifies any positive influence FPI
has on
growth. Indeed, this positive influence is more rapidly negated than the
influence on
capital issuance. This is not particularly surprising given the risk aversion
during crisis
periods and the reactions firms have with regard to their operations.
[Insert Table XII here]
B.2 Out-of-sample
Using estimates of propensity for crisis outside of the sample period instead
of
within, I examine the cross-section of the sample in each year, utilize the
fitted value of
equation (6) for the preceding four-year period (i.e. 1991-1995 for
time=1996, 1992-1996
for time=1997, etc.). Doing so provides more detail in the results, which
highlights the
39 The inclusion of this variable also may bias upward a country’s propensity
for risis making the actual
59
Asian Crisis and its contagion in the results. Years other than 1998-99 offer
very similar
results to those in the in-sample analysis. The two-year period of the crisis
interrupts
these relationships quite a bit. The interactive term looses its significance,
more than
likely because a significant portion of the sample is either in crisis or
influenced by crisis
due to contagion. Fitting in nicely with this is the fact that we see that the
coefficient for
FPI is actually negative here. FPI provides value as long as a country’s
propensity for
crisis is not above the average for the sample – the significant difference
between the
variables of interest as well as the majority of the control variables
demonstrates nicely
how few macroeconomic factors aren’t affected negatively by crisis making
the case that
FPI is one factor among many that may lead to decreased access to capital
when a
country is in crisis. These results may be seen in Table XIII.
[Insert Table XIII here]
VI. Conclusions
Although foreign portfolio investment serves a potential additional source of
investment capital for small firms, the volatility of this capital flow in times of
crisis
threatens the benefits FPI provides such as an enhanced access to capital.
Importantly,
the short-term growth of these firms seems to be relatively unaffected by the
variability in
this capital flow, except in those periods of decreased investor confidence
(alternatively -
in periods of higher country risk). In these less “investible” periods, FPI
volatility
hinders the small firm when taking growth into consideration, implying that
access to
finance may be interrupted, and that the risk aversion that ensues with
volatility in these
point at which FPI’s benefits are neutralizes higher than 24%.
60
capital flows decreases benefits derived from it in these times through
decreased
liquidity.
Results in this paper support the contention that volatility of capital flows is
potential damaging to host economies. Specifically, FPI volatility can
interrupt enhanced
access to finance for small firms through a reduction in the liquidity of
financial markets.
The results do not support the contention that volatility is harmful in all
times, finding
that waves of investment do not significantly decrease the probability that a
small firm is
able to issue capital in the public markets in times when investor confidence
is increased
and does not necessarily hinder firm growth in the short term. A policy
implication of
this is that countries should try to stabilize capital flows by way of increasing
institutional
investor confidence in their nation. Fortunately, having open borders to
foreign investors
goes part of the way toward that end, since liberalized nations see increases
in both the
size and the liquidity of markets, as well as improvements in corporate
governance and
disclosure.
61
Dissertation Conclusions
Foreign portfolio investment has the potential to influence foreign
investment at
the country, firm and investor level to motivate actions which influence
change to better
the investment environment, which will in turn draw more investment. The
longer-term
benefits of this capital flow such as improvements in corporate governance,
and
disclosure are not likely to reverse based on fluctuations in this flow. Shorter-
term
benefits of this capital flow, such as market liquidity, may reverse in the
presence of
severe fluctuations. These fluctuations and the resulting decrease in liquidity
may serve
to lessen or negate the benefits of the capital flow itself. Results of
examinations of both
the capital flow and its volatility imply that although the benefits of foreign
portfolio
investment may be potentially economically large and worthy of market
integration, the
instability has the potential to destroy the benefits initially derived by the
capital flow.
Based on the results of this study, policy implications are that countries that
wish
to benefit from foreign portfolio investment should strive to improve property
rights and
investor protection such that confidence in these countries reaches a level
which is
minimum for investors to remain when macroeconomic changes occur. In the
end,
policies that will minimize volatility while still allowing for the benefits of
foreign
portfolio investment may be put into place to begin a beneficial investment
cycle that will
improve foreign investment and domestic financial development for the long
run.